Traditional methods of creating and displaying terrain on computers have relied on creating a database that fits completely into main memory. This involves loading geospecific terrain height data and image data at the start of the program. This implementation becomes limited when databases of sufficient resolution and size increase to surpass system memory.
In recognition of this problem, data loading advancements typically involve the “paging in” of data as needed. However, the techniques used to accomplish this and the speed at which they can retrieve and display the data are where image processing software systems differ. Speed is very important in the viewing of visual terrain environments when one wants to quickly move form one area of the world to another. Furthermore, one frequently wants to “fly through” the database. That is, the data may be initially viewed from great altitudes above the terrain while subsequent viewing might be required at a much lower altitude. Since processing memory at any given time is limited, not all data may be loaded when needed. Thus, data associated with a given area/location must be readily identifiable/locatable in the storage database.